Key takeaways
- The AI Omnibus, Regulation (EU) 2026/1744, is adopted law: published in the Official Journal on 24 July 2026 and in force since 27 July 2026.
- Annex III high-risk rules, which cover AI that evaluates staff or allocates their work, apply from 2 December 2027.
- Banned practices and the transparency rules for AI that talks to people already apply, so 2027 is not the first deadline.
- Most prospecting, research and writing uses sit outside the high-risk list; judge each tool by how it is used, not by its label.
- This page is general information, not legal advice; take your own counsel on how the Act applies to your business.
What changes for sales teams under the EU AI Act in 2027
The change in 2027 is that high-risk duties begin for the AI uses listed in Annex III, from 2 December 2027. For a sales organisation the listed area that matters most is employment: AI intended to make decisions on promotion or termination, to allocate tasks based on individual behaviour or personal traits, or to monitor and evaluate how staff perform and behave.
That reaches tools many revenue teams run on their own people: rep scorecards produced by AI, coaching systems that rate each person's calls, and engines that hand out accounts or leads according to how each rep behaves. A provider can document that a listed system poses no significant risk, for example because it only performs a narrow procedural task, but an Annex III system that profiles people is always treated as high-risk.
The date itself was moved by the AI Omnibus. The Act as first adopted applied the Annex III rules from 2 August 2026; Regulation (EU) 2026/1744 set 2 December 2027 instead, and 2 August 2028 for AI built into regulated products. It was published in the Official Journal on 24 July 2026 and is law, not a proposal.
What does a deployer of high-risk AI owe from December 2027
A company that uses a high-risk system under its own authority is its deployer, and from 2 December 2027 it carries duties of its own under Article 26, beside the provider's. For a sales leader they are practical rather than technical.
- Use the system as the provider's instructions for use describe, and keep any input data you control relevant to its purpose.
- Assign human oversight to people with the competence, training and authority to act on what the system says.
- Keep the logs the system generates for at least six months, unless another law sets a different period.
- As an employer, tell workers' representatives and the affected reps before the system is used at work.
- Monitor how it performs, and stop using it and tell the provider if it presents a risk.
Which EU AI Act rules already apply to sales before 2027
Two sets of rules already apply, so a team waiting for 2027 is late on both. Since 2 February 2025 the Act has banned some practices outright, among them AI that infers the emotions of people at work, outside medical or safety reasons, and AI that uses manipulative or deceptive techniques to distort someone's decision in a way that causes significant harm.
Since 2 August 2026 the Article 50 transparency rules apply. A provider of an AI system meant to interact directly with people, such as a chat or voice agent, must design it so that people know they are dealing with AI unless that is obvious, and a deployer that publishes deepfake audio or video must disclose that it was generated or manipulated. Providers of generative tools already on the market before 2 August 2026 have until 2 December 2026 to mark their output in a machine-readable way.
If your team builds its own agent and puts it into service under your company's name, you are its provider for these purposes, not only its user.
How should a sales team prepare for each AI Act date
Prepare in date order: first the duties already in force, then the marking deadline, then the high-risk list.
- Now: list every AI tool the team uses, what it decides or produces, and whether you are its provider or its deployer.
- Now: switch off any feature that reads your own reps' emotions, and check that chat and voice agents tell people they are AI.
- Now: give the people who use AI the guidance and training they need, which the AI literacy article asks providers and deployers to support.
- Before 2 December 2026: ask the vendors of your generative tools how their output is marked.
- Before 2 December 2027: find any system that scores, ranks, monitors or assigns work to reps, ask its provider whether it is high-risk, and plan the oversight, logs and notice to staff.
What stays the same for sales teams in 2027
Most of the selling job stays off the AI Act's high-risk list in 2027. Finding companies, researching accounts, drafting emails a person reviews and keeping a CRM current are not listed uses, so for them the duties are the general ones: no banned practices, transparency where it applies and support for AI literacy.
GDPR still governs the personal data in every prospect list, and national e-marketing rules still decide whom you may email without consent; the AI Act sits beside both rather than replacing them. Your company also still answers for what it sends, whichever tool wrote the draft.
This page is general information, not legal advice. How the Act applies depends on the system, its intended purpose and your role, so take advice from your own counsel before relying on any reading here.
EU AI Act dates for B2B sales and marketing teams
The application dates as they stand after the AI Omnibus, with what each one brings and whom it reaches in a commercial team.
| What applies | Who it affects | |
|---|---|---|
| 2 February 2025 | Banned practices, including AI that infers emotions at work and manipulative techniques that distort decisions; the AI literacy article | Every provider and deployer, including sales teams whose call tools read their own reps' emotions |
| 2 August 2025 | Obligations for providers of general-purpose AI models, and the governance rules | Model providers; sales teams only through what their vendors disclose |
| 2 August 2026 | Most of the Act applies, including the Article 50 transparency rules, and enforcement starts for the rules in force | Providers of chat and voice agents that talk to people; deployers that publish deepfake audio or video |
| 2 December 2026 | Two new prohibitions added by the AI Omnibus; generative tools on the market before 2 August 2026 must meet the Article 50(2) marking rule | Providers of generative AI; sales and marketing teams check their vendors |
| 2 December 2027 | High-risk rules for Annex III systems, including employment uses such as monitoring and evaluating staff or allocating tasks by individual behaviour | Providers and deployers of those systems, including employers that use AI to score, rank or assign work to reps |
| 2 August 2028 | High-risk rules for AI built into products covered by Annex I | Product manufacturers; rarely a sales team |
Sources: Article 113 of Regulation (EU) 2024/1689 as amended by Regulation (EU) 2026/1744 (the AI Omnibus, Official Journal of 24 July 2026, in force since 27 July 2026), with Articles 2, 4, 5, 6, 26, 50 and 111 and Annex III, on EUR-Lex and the European Commission's AI Act Service Desk, checked on 13 November 2026. General information, not legal advice.
Where Pebbles Ai fits
- CapabilityAI workspacePebbles Ai is an AI workspace whose assistants research accounts with cited sources, find fresh leads with work emails and write outreach and social posts from your company knowledge, and its execution mode gives your AI inventory a named review point: in Manual, sends, posts, CRM writes and MCP calls wait for your approval, in Plan they are queued and in Auto they run without asking. It is a control you set, not a certification of compliance.
- FoundationAgent memoryPebbles Ai learns your company as you work through Company Intel and Agent memory, and no customer's data trains a model. Saved memories are listed, searched and forgotten in Settings, where only admins forget company memories. Tenant-Isolated AI, the Pebbles Ai architecture for tenant isolation and data security, keeps each business in its own secured tenant; Pebbles Ai is GDPR compliant and a DPA is available.
Common questions
Did the AI Omnibus delay the EU AI Act high-risk deadline to 2027?
Yes. Regulation (EU) 2026/1744 was published in the Official Journal on 24 July 2026 and entered into force on 27 July 2026. It moved the Annex III high-risk rules to 2 December 2027 and the rules for AI in regulated products to 2 August 2028. It is in force, not a proposal.
Does the EU AI Act apply to a US sales team selling into Europe in 2027?
It can. Article 2 covers providers placing AI systems on the EU market wherever they are based, and providers and deployers outside the EU where the system's output is used in the EU. Check where each tool's output is used, not only where your team sits.
Is AI lead scoring high-risk under the EU AI Act in 2027?
Scoring companies or prospects for sales priority is not a use listed in Annex III, so it is not high-risk on that ground. Scoring your own reps, or allocating leads by each rep's behaviour, falls under the employment uses and needs a closer look before 2 December 2027.
Can sales teams keep using AI call coaching under the EU AI Act in 2027?
Yes, within limits. A feature that infers your reps' emotions at work has been banned since 2 February 2025, outside medical or safety reasons. Coaching that monitors and evaluates each rep's performance is a listed high-risk use from 2 December 2027, so ask the provider how it classifies the system.
About this page. Written by the Pebbles Ai team for people who run sales, marketing and growth. Last updated .